On which corpus was BartTokenizer trained?

#8
by boydcheung - opened

Could you specify the corpus and if I would like to extend the vocabulary, what is the expected way?

Sign up or log in to comment